JP6425735B2 - Determination of temporary trade limits - Google Patents

Description

[Cross-reference to related technologies]
Priority is claimed on Chinese Patent Application No. 20141011385.2, filed on March 25, 2014, and entitled "A METHOD AND A PLATFORM FOR PROCESSING BIG DATA". Claim the right. This application is incorporated herein by reference for all purposes.

  The present application relates to the field of big data processing technology, and in particular to technology for processing big data.

  The Internet generates huge amounts of data daily. For example, blogs, mini-blogs, and trading platforms such as Twitter, Facebook, and other social networks are generating large amounts of data daily. Data is pervading every industrial and industrial function. In some businesses, transaction process, product usage, and human behavior information are all converted into data.

  Although this generated data appears to be unpopular, unconsolidated, and does not follow any rules that are easily identifiable, it does have certain characteristics that can reflect certain characteristics in terms of its overall distribution Have. The issue of how to mine and process this huge and messy data is an important research theme in the field of big data and data mining. Data mining is the process of extracting information and knowledge from incomplete, noisy, ambiguous, and / or random, large amounts of actual application data. The exact type of knowledge that will be included in such data will not be known to people before that data mining, so the knowledge that may be mined from the data may be very useful There is.

  An important type of processing in big data mining is the clustering procedure. A clustering procedure can divide a large set of data objects into a series of significant subsets, or clusters. Cluster analysis consists of dividing groups of data objects into several categories according to similarities and differences. The goal of cluster analysis is that it results in the highest similarity between data belonging to the same category and the lowest similarity between data of different categories. Cluster analysis can be applied to customer classification, analysis of customer background, prediction of customer buying habits, market segmentation, and other areas.

  Cluster analysis generally involves grouping of data objects, where similar data objects are grouped into one category. A typical clustering method is k-means clustering. The k-means clustering technique receives inputs of k clusters and receives a database containing N data objects. The k-means clustering technique takes these N data objects and outputs them into k clusters that meet the least squares criterion. Among N data objects allocated to k clusters, data objects in the same cluster have high similarity to one another, and data objects in different clusters have low similarity to one another. Overall, this type of cluster similarity can be calculated using the "center object" (attractive center) obtained from the average of data objects in each cluster.

  The process of implementing the k-means clustering technique specifically includes:

  (1) Select any k objects as initial cluster centers from N data objects.

  (2) Using the average of the objects of each cluster (center objects), calculate the distances from each object to these center objects, and subdivide the corresponding objects according to the minimum distance.

  Recalculate the average (central object) of each cluster (if any).

  (4) Calculate a standard measurement function. The clustering technique ends when certain conditions are met, eg, the function converges. If the condition is not met, the clustering technique returns to step (2).

  However, conventional clustering analysis suffers from at least some drawbacks. The first drawback is that the k-means clustering technique requires that the input number N of objects be a fixed value. In a situation where the value of N is a variable value, for example, every time N changes during processing, such as increasing the value of N, it is necessary to perform processing for adding one new data record, , Re-execution of steps (1) to (4) as described above is required. The second drawback is that when processing high quality data, a significant amount of hardware resources (eg, memory and processor resources) are required to perform clustering as described above is there.

  Various embodiments of the present invention are disclosed in the following detailed description and the accompanying drawings.

FIG. 1 illustrates an embodiment of a system for determining temporary transaction limits for a user's transaction.

FIG. 7 is a flow chart illustrating an embodiment of a process for determining temporary transaction limits for a user's transaction.

FIG. 5 is a flow chart illustrating an example of a process for determining a user's predetermined limits.

FIG. 6 is a diagram showing an example of data points in a group.

From the data point d ₁ in the population is a diagram showing an example of a d _7.

FIG. 6 is a diagram showing an example of a portion of a decision tree that classifies users based on user RFM values.

FIG. 1 illustrates an embodiment of a system for determining temporary transaction limits for a user's transaction.

FIG. 5 illustrates an embodiment of a predetermined limit calculation unit.

FIG. 1 is a functional diagram illustrating one embodiment of a programmed computer system for performing temporary transaction restriction generation.

  The invention relates to a process, an apparatus, a system, a composition, a computer program product implemented on a computer readable storage medium, and / or instructions stored in a memory to which it is connected and / or instructions provided by memory to which it is connected. Can be implemented in numerous forms, including a processor such as a processor configured to perform In this specification, these implementations, or any other form that the present invention may take, may be referred to as techniques. Overall, the order of the disclosed process steps may be varied within the scope of the present invention. Unless otherwise stated, a component such as a processor or memory described as being configured to perform a task performs the task as a general purpose component temporarily configured to perform the task at a given time, or It may be implemented as a specialized component manufactured as: The term "processor" as used herein refers to one or more devices, circuits, and / or processing cores configured to process data, such as computer program instructions.

  A detailed description of one or more embodiments of the present invention is provided below, along with accompanying figures that illustrate the principles of the present invention. The invention is described in connection with such embodiments, but is not limited to any of the embodiments. The scope of the present invention is limited only by the claims, and the present invention includes numerous alternatives, modifications and equivalents. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. These details are provided for the purpose of illustration, and the present invention may be practiced according to the claims without some or all of these details. For the purpose of clarity, technical material that is known in the technical fields related to the present invention has not been described in detail so that the present invention is not unnecessarily obscured.

  As e-commerce becomes increasingly popular, payment platforms are playing an ever more important role in people's daily lives. By using a credit card issued to the user, the payment platform has made the payment tremendously convenient for the user and has created an excellent payment experience for the user. However, improving the convenience of payment has also increased the risk of transactions, and has imposed even more stringent requirements on risk control. One important means of risk control is the limitations imposed on the user's consumption. The payment platform can impose limit management on user transactions. A relatively common approach is to limit trades according to fixed limits. For example, a whitelist can be set up for the payment platform. Users who are whitelisted can be given higher limits, while users who are not whitelisted can be given lower limits. In this way, the user's consumption can be constrained based on the assessment of whether the consumption presented by the user exceeds the user's limit. In such an approach, the user's consumption limit is usually a manually set fixed value. Stated differently, in the prior art, the restrictions imposed on the consumption of the user are fixed and can not be adjusted on a per-user and per-situ basis.

  Described herein are embodiments for determining temporary transaction limits. A transaction request message is received. The transaction request message includes a request for approval of the transaction. The transaction request message includes at least a total transaction cost and identification information associated with the requesting user. If the total transaction cost exceeds a predetermined limit associated with the user, past transaction data associated with various users is retrieved and analyzed to form a clustered classification. The clustered classification corresponding to the user associated with the transaction request message is determined using past transaction data associated with the user. Predefined mapping rules are used to determine the dynamic quota corresponding to the clustered classification corresponding to the user. Whether the transaction request message is approved is determined based on comparing the total transaction cost to the combination of the predetermined limit and the dynamic quota.

  FIG. 1 is an illustration of an embodiment of a system for determining temporary transaction limits for a user's transaction. In this example, system 100 includes device 102, network 104, electronic commerce platform server 106, and trading platform server 108.

  Devices 102 include devices through which users can access electronic commerce platforms associated with electronic commerce platform server 106. For example, device 102 may include a smartphone, a tablet terminal, a desktop computer, a laptop device, a mobile phone, and / or any type of computing device. The e-commerce platform provided by e-commerce platform server 106 may be configured to display a user interface associated with the display of profiles / information associated with various products to a website and / or an application running on device 102. Will present.

  The user can use the device 102 to view and select products for purchase from an e-commerce platform. Once the user has selected at the e-commerce platform one or more products that he / she wants to buy, the e-commerce platform is able to trade in the identification information associated with the user and the total cost of the trade (total transaction cost), etc. A transaction request message can be generated that includes the associated information. The e-commerce platform server 106 is configured to send this transaction request message to the transaction platform server 108.

  Trading platform server 108 is configured to determine whether the transaction request message should be approved. Trading platform server 108 is configured to determine whether the total transaction cost exceeds a predetermined limit associated with the user. In various implementations, the trading platform server 108 is configured to retrieve past transaction data associated with various users and analyze the retrieved data to determine a clustered classification. . Trading platform server 108 is configured to determine the clustered classification corresponding to the user using past transaction data associated with the user. Trading platform server 108 is configured to determine a dynamic quota corresponding to the user based on the clustered classification corresponding to the user and the predetermined mapping rules. Trading platform server 108 is configured to determine whether the transaction request message is approved based at least in part on comparing the total transaction cost to a combination of predetermined limits and dynamic quotas. Be done.

  Thus, in various implementations, data mining may be used to flexibly adjust trade limits on a per-user basis among various situations.

  FIG. 2 is a flow chart illustrating an embodiment of a process for determining temporary transaction limits for a user's transaction. In some embodiments, process 200 is implemented in system 100 of FIG.

  At 202, a transaction request message is received. The transaction request message includes the total transaction cost and identification information associated with the user.

  The transaction request message includes a request for the user to approve the pending transaction. The transaction request message includes information describing the pending order of the user. For example, the user has selected one or more products to purchase from the e-commerce platform. For example, the type of product to be purchased, the cost of the individual product to be purchased, the total cost of the product to be purchased (total transaction cost) so that the e-commerce platform server can determine whether the transaction can be approved And the identifier of the user who placed the order may be included in the transaction request message by the electronic commerce platform server.

  At 204, it is determined that the total transaction cost exceeds the predetermined limit.

  In some embodiments, predetermined per-transaction limits for the user may be set manually, for example, by a system administrator, by the user, or by default.

  In some other embodiments, the predetermined limit per transaction for the user may be determined dynamically by the payment platform. For example, the payment platform may calculate predetermined limits according to specific rules. In one implementation, the payment platform may use a density based outlier exclusion model to determine predetermined limits. FIG. 3, described below, illustrates an example in which predetermined limits are determined for the user using a density based outlier exclusion model.

  If it is determined that the total transaction cost included in the transaction request message is less than or equal to a predetermined limit per transaction for the user, then the transaction associated with the transaction request message is approved and the process 200 ends. Do. Rather, if it is determined that the total transaction cost included in the transaction request message exceeds the predetermined limit for each transaction for the user, then the transaction associated with the transaction request message may be determined by the user. To determine the amount (dynamic quota) to increase (temporarily) the limit of E., using the steps described below.

  At 206, upon determination that the total transaction cost exceeds the predetermined limit, historical transaction data associated with the plurality of users is retrieved.

  The past associated with various users, including the user associated with the current transaction request message, after it is determined that the total transaction cost included in the transaction request message exceeds the predetermined limit per transaction for the user Transaction data is retrieved from storage. The various users who have retrieved past transaction data may include users of e-commerce platforms and / or payment platforms that have previously issued transaction request messages. The past transaction data associated with each user may be, for example, when the user made a purchase (eg, a timestamp associated with a past transaction), what the user purchased (eg, each past transaction) Products and / or other items involved, as well as data describing expenses associated with the user's past purchases (eg, how much the user has spent in each past transaction).

  In some embodiments, past transaction data of users associated with the trading platform and / or e-commerce platform may be stored in a data warehouse. For example, a data warehouse contains a huge database. Such a database stores integrated data from a task organized database. For example, a data warehouse stores data obtained from a transaction processing system (eg, online transaction processing (OLTP)). However, the process of transforming and integrating data can be challenging when establishing a data warehouse. The goal for a data warehouse is to convert activity data into useful strategic information. In some embodiments, the data warehouse stores the following types of data: consolidated data, detailed and summarized data, historical data, and data for data interpretation.

  At 208, a plurality of clustered classifications are determined from past transaction data associated with a plurality of users.

  Data mining is applied to retrieved past transaction data associated with various users to generate one or more clusters. Here, each cluster is associated with a corresponding classification. Thus, in some embodiments, a cluster may be referred to as a "clustered classification."

  The main purpose of establishing a data warehouse to store past transaction data associated with various users, and applying data mining to the data warehouse, is useful for making decisions from the data warehouse Mining information and knowledge. Usually, a data warehouse needs to be established before data mining can proceed efficiently. The data contained in the data warehouse should be clean itself (eg, data stored in the data warehouse should preferably not be corrupted), complete and integrated desirable. Thus, the relationship between establishing a data warehouse and data mining there from will probably be understood as follows. That is, data mining is a process and technique for finding useful information in a huge data warehouse. In most situations, data mining involves moving data from a data warehouse to a data mining database or data store. In order to convert a large amount of data into useful information, it is first necessary to collect information efficiently. A database system with a complete set of features has become a preferred tool for collecting data. Simply put, data warehousing is to collect useful data from other systems and put them into an integrated storage area. In some embodiments, a specially large relational database, which has been processed and integrated, to store data needed by the decision support system, such as data used for decision support or data analysis. used. Raw data can be structured, for example, as data in a relational database. Raw data can also be semi-structured as text, graphics and image data. Raw data can, in turn, be disparate data distributed over the network. Any suitable data discovery technique is applicable.

  In some embodiments, clustering data mining techniques are applied to a predetermined amount of past transaction data associated with a user to determine one or more clustered classifications. The predetermined amount may be configured to be a sample of data objects that properly represents the entire past transaction data.

  Data mining differs from traditional data analysis (e.g., query or statement analysis or online analysis processing) in that it digs out information and discovers knowledge without any clear cut assumptions. It is desirable that the information obtained through data mining have three characteristics: previously unknown, effective, and practical. Information that has not been known until now is information that was not necessarily predicted in advance. Data mining is desirable to discover information or knowledge that may be intuitively discovered, or to discover information or knowledge that is counterintuitive.

  In some embodiments, past transaction data stored for each user may include one or more data objects associated with one or more transaction data. In one implementation, each user data object may include data on three metrics: the latest consumption, the frequency of consumption, and the amount of money consumed.

  REM (final purchase date, purchase frequency, and purchase price, referring to the latest consumption, consumption frequency, and consumption price, respectively) models are widely used in many customer relationship management (CRM) analysis models. It is used. RFM models are considered to be important tools and tools to measure customer value and customer fortune capability. This model describes the customer's value state in terms of three measures: the latest purchase activity, the total frequency of purchases, and the amount of money consumed. Logically, the customer close to the last consumption time is the customer with high possibility of consumption. These customers are the ones most likely to respond to the latest product or service offering. Large amounts of historical data indicate that customers who have already purchased are very likely to continue purchasing. The frequency of consumption is the number of purchases made by the customer within a set period of time. It can be said that the most frequently purchased customers are the most satisfied customers. If the customer believes in the brand or store's commitment, the customer who purchases the most often will be the one with the highest adherence. The consumption amount may be the total amount of consumption within a set period. Consumption is the most effective predictor. According to Pareto's law, 80% of a company's revenue comes from its 20% of its customers. The amount consumed by the top 10% of customers is considered to be at least twice the amount of consumption by customers in the next tier, and will account for more than 40% of the company's total business. In terms of cumulative rates, it can be seen that 40% of customers contribute 80% of the company's total turnover and 60% account for over 90% of the enterprise's turnover.

  In some embodiments, predetermined amounts of past transaction data associated with different users are clustered in consideration of three measures: current consumption, frequency of consumption, and consumption amount. In this way, each resulting clustered classification will be associated with a range of values associated with the latest consumption, frequency of consumption, and consumption metrics. For example, according to the results of the clusters, three clustering classifications may be formed, the first clustering classification being associated with high (e.g. the highest range of values for each measure) and the second clustering classification The classifications are associated with medium (eg, the second highest range of values for each scale) and the third clustering classification is associated with low (eg, the lowest range of values for each scale). For example, clustering of historical transaction data may be performed using any suitable clustering technique, such as k-means clustering techniques, density-based spatial clustering (DBSCAN) techniques for noisy applications, or hierarchical clustering techniques. .

  In some embodiments, steps 208 and 210 need to be performed for each subsequent transaction request message if these steps have been performed recently (eg, within a recent predetermined time period) There is no. Stated differently, clustered classifications are not necessarily required each time a transaction request message is received, but can be determined periodically.

  At 210, a clustered classification of a plurality of clustered classifications is determined for the user based at least in part on past transaction data associated with the user.

  Once the clustering of past transaction data associated with various users is complete and the corresponding classification obtained, this is done for that user based on the user's historical data associated with the transaction request message. One clustering classification can be determined. For example, the three clustered classifications created as mentioned above are represented by high, medium and low.

  In various embodiments, the user's past data associated with the transaction request message is converted into numerical values (R value, F value, and M value) using the RFM method. The stored one or more decision trees may then be used in conjunction with RFM values derived from the user's past data to determine the appropriate one clustered classification for the requesting user. For example, to obtain a user's RFM value, scoring can be performed according to stored RFM scoring rules using the user's past data.

  FIG. 6 illustrates an example of a portion of a decision tree that classifies users based on their RFM values. An example of the part of the decision tree shown in FIG. 6 is that users whose RFM values simultaneously satisfy R> 3.5, F> 4 and M> 4 are classified as users of “high” clustered classification Is shown. Another part of the decision tree not shown in FIG. 6 is, for example, that users with RFM values simultaneously satisfying R <2, F <2, and M <2 are classified into the “low” clustered classification, and It may indicate that the remaining users considered not to be classified as "high" clustering classification or "low" clustering classification by the RFM value can be classified as users of "medium" clustering classification.

  For example, the historical data of the user associated with the transaction request message is converted to the following RFM values: R = 3.7, F = 4.1, and M = 4.3. Thus, based on the portion of the decision tree shown in FIG. 6, users with R = 3.7, F = 4.1 and M = 4.3 may be classified as “high” clustered classifications.

  Returning to step 212, using the predetermined mapping rules, a dynamic quota corresponding to the clustered classification for the user is determined.

  In various embodiments, the "dynamic quota" includes a maximum value above a predetermined limit that can increase the total transaction cost of the user so far to cause the request transaction to be approved. In various embodiments, predetermined mapping rules are stored that associate each clustered classification (e.g., high, medium and low) with corresponding dynamic quota values. For example, "high" clustering classification is mapped to $ 2,000 dynamic quota through mapping rules, "medium" clustering classification is mapped to $ 1,000 dynamic quota through mapping rules, "low" 'Clustered classifications are mapped to $ 500 dynamic quotas.

  At 214, it is determined whether the transaction request message is approved based at least in part on comparing the total transaction cost to the temporary transaction limit. Here, temporary trade restrictions include a combination of predetermined restrictions and dynamic quotas.

  Prior to this, at 204, it has been determined that the total transaction cost included in the transaction request message exceeds the predetermined limit associated with the user, so that the user's trade limit is at least partially clustered. According to the analysis result, it is determined that the dynamic quota determined temporarily based on the clustered classification associated with the user is temporarily raised. In various embodiments, the temporary transaction limit is the sum of a predetermined limit and a dynamic quota. Since the dynamic quota can be determined for each transaction of the user, whereas the predetermined limit for the user does not change through multiple transactions, the dynamic quota determined for the user is It can be understood as raising the predetermined limit temporarily. If the total transaction cost is less than or equal to the predetermined limit plus the dynamic quota, the transaction associated with the transaction request message is approved. Otherwise, if the total transaction cost exceeds the predetermined limit plus the dynamic quota, the transaction associated with the transaction request message is not approved.

  In some embodiments, clustering of a predetermined amount of past transaction data may be completed prior to process 200 and / or may not be repeated each time process 200. When a new transaction request message is received from the user, it is not necessary to recluster a large amount of data including this newly received data. On the contrary, among the classifications determined so far based on the past transaction data associated with the user associated with the transaction request message, the corresponding clustered classification is searched, and according to the predetermined mapping rule, You can get a dynamic quota. Thus, the sum of the predetermined limit and the dynamic quota may act as a temporary transaction limit for the user, and then the total transaction cost included in the transaction request message sent by the user may be approved. It is possible to make decisions based on temporary trade restrictions.

  The following is an example of applying the process 200. A trading request message is received at the trading platform server that includes identification information associated with user A. The total transaction cost included in this request message is $ 2,500. The predetermined limit for this user is, for example, $ 2,000. In some embodiments, the predetermined limit is a fixed value. In some embodiments, the predetermined limits are determined dynamically using the techniques described below in connection with FIG. The clustering classification obtained for this user according to the user's past data from among a plurality of cluster classifications is in, and according to a predetermined mapping rule, the dynamic quota corresponding to the clustering classification in is It is $ 1,000. Thus, the temporary limit for the user is the sum of the predetermined limit and the dynamic quota, ie $ 1,000 + $ 2,000 = $ 3,000. The total transaction cost $ 2,500 associated with the transaction request message is less than the temporary limit $ 3,000. Thus, the transaction associated with the transaction request message is acceptable. For example, if the transaction request message is approved, then the request transaction associated with the total transaction cost can be completed, and the product associated with the transaction request message is approved to be purchased by the requesting user Be done. However, if the total transaction cost associated with the transaction request message is $ 3,500, this would exceed the temporary limit of $ 3,000 and be associated with the transaction request message. Trades will be rejected.

  FIG. 3 is a flow chart illustrating an example of a process for determining a user's predetermined limits. In some embodiments, process 300 is implemented at system 100 of FIG. In some embodiments, the predetermined limits used in step 204 of process 200 of FIG. 2 may be determined using the example of process 300.

  In density-based outlier exclusion models, the difference between data points is considered to be a "distance" in one-dimensional space, and the entire data point is considered to be a "population." Thus, data points that are specifically interspersed within each "population" are considered to be "outliers". In the example of process 300, the data points include the user's past transaction costs. The user's past transaction costs include the total costs associated with the past transactions requested by the user. As shown in the example of FIG. 4, the data points in region D are relatively concentrated and are high density points in the population. The data points in the regions E, F, G are relatively scattered and thus referred to as "outliers". Process 300 describes an example where outlier past transaction costs associated with a user are excluded and not used to determine predetermined limits for the user.

  At 302, a plurality of past transaction costs associated with the user are ranked.

In some embodiments, n past transaction costs associated with the user may be ranked based on the transaction costs. For example, transaction costs are ranked in order from the smallest. For example, ranked past transaction costs are d ₁ , d ₂ ,..., D _n .

  At 304, the outlier condition is determined using a predetermined equation.

  In some embodiments, the predetermined equation is:

In equation (1), i denotes the ith transaction, d ₁ , d ₂ ,..., D _n is the ranked past transaction costs, C is set to a threshold, and k is Set to a preset distance. Therefore, according to the above described equation, when the distance from the i th transaction costs (d _i) to k transaction Bundake apart transaction costs exceeds the threshold C, i th transaction costs (d _i) is disengaged It is considered to be a value.

  Thus, by setting the values C and k, equation (1) can be used to determine the ith (one or more) past transaction costs that satisfy the condition to be an outlier .

  At 306, among the plurality of past transaction costs, one or more past transaction costs that meet the outlier condition are determined.

As an example using a specific number, one may refer to FIG. FIG. 5 shows data points associated with the past transaction costs groups d ₁ -d ₇ . The past transaction costs d _{1 to} d ₇ are listed in ascending order of d ₁ = 100, d ₂ = 110, d ₃ = 123, d ₄ = 195, d ₅ = 229, d ₆ = 1410, and d ₇ = 2100. is there. In the example of FIG. 5, C is set to 300 and k is set to 3. Therefore, by applying equation (1), the following is obtained.

Thus, d ₆ and d ₇ are determined to be outliers and excluded from the population. The remaining data points in the population include d _{1 to} d ₅ .

  At 308, a predetermined limit associated with the user is determined based at least in part on a maximum value from the plurality of past transaction costs excluding one or more past transaction costs that satisfy the outlier condition.

The maximum value of d _{1 to} d ₅ is 229. Thus, the maximum value 229 is set as the predetermined limit for the user.

  In some embodiments, the transaction request message may include one or more product types associated with the transaction. That is, the transaction request message may include the individual costs of each of a plurality of different types of products. These multiple different product types may correspond to different trading situations. If there are multiple transaction situations associated with the transaction request message, each such situation may be associated with a predetermined limit.

  FIG. 7 is an illustration of an embodiment of a system for determining temporary transaction limits for a user's transaction. In this example, system 700 includes a receiving unit 710, a querying unit 720, a searching unit 730, a mapping unit 740, and a verification unit 750.

  These units may be implemented as software components executing on one or more general purpose processors, or as hardware such as programmable logic devices and / or designed special purpose integrated circuits. These elements include some instructions for causing a computer device (such as a personal computer, a server, a network device, etc.) to execute the method described in the embodiment of the present invention, and (optical disk, flash storage device, mobile) It can be embodied in the form of a software product that can be stored on a non-volatile storage medium, such as a hard disk. These units may be implemented on one device or distributed to multiple devices.

  The receiving unit 710 is configured to receive the transaction request message. The transaction request message includes the total transaction cost and identification information associated with the user. The transaction request message may describe one or more clusters of product types associated with the transaction.

  The inquiry unit 720 is configured to compare the total transaction costs with predetermined limits associated with the user. The inquiry unit 720 is configured to inquire about past transaction data associated with various users if the total transaction cost associated with the transaction request message received by the receiving unit 710 exceeds a predetermined limit. Ru.

  The search unit 730 is configured to determine a plurality of clustered classifications from at least a portion of past transaction data associated with various users. In some embodiments, the clustered classification takes into account three measures, namely the latest consumption, the frequency of consumption, and the amount of consumption, for a predetermined amount of past transaction data associated with different users. Obtained after applying the clustering procedure. By using big data simulation scenarios, new needs are identified and the return on investment is increased. Big data analysis technology enables vendors to store and analyze this data in real time in conjunction with traditional behavioral data, and to do it more cost effectively. With big data technology, this data can be integrated and go through data mining. In doing so, model simulations can be used in some circumstances to assess which approach has the highest return on investment under various variables.

  The search unit 730 is also configured to determine a clustered classification corresponding to the user associated with the transaction request message based on past transaction data of the user associated with the transaction request message.

  The mapping unit 740 is configured to determine the dynamic quota corresponding to the clustered classification associated with the user associated with the transaction request message according to a predetermined mapping rule.

  Verification unit 750 is configured to approve the transaction request message if the total transaction cost is less than or equal to the predetermined limit plus the dynamic quota. Verification unit 750 is configured to reject the transaction request message if the total transaction cost exceeds the predetermined limit plus the dynamic quota.

  FIG. 8 is a diagram showing an embodiment of the predetermined limit calculation unit. In this example, predetermined limit calculation unit 860 of system 800 is connected to receiving unit 810 and querying unit 820. The receiving unit 810, the querying unit 820, the searching unit 830, the mapping unit 840 and the verifying unit 850 are implemented similarly to the receiving unit 710, the querying unit 720, the search unit 730, the mapping unit 740 and the verifying unit 750 of FIG. be able to.

  The ranking unit 861 is configured to rank past transaction costs associated with the user. For example, the user is associated with the transaction request message.

  The assessment unit 862 is configured to determine the outlier condition using a predetermined equation.

Equation (1) is that, as described above, i represents the i-th transaction, d ₁ , d ₂ ,..., D _n are the past transaction costs ranked, and C is set to a threshold, k Is a predetermined equation set to the preset distance. Thus, using the equation described above, when the distance from the i th transaction cost (d _i ) to the transaction cost separated by k transactions exceeds the threshold C, the i th transaction cost (d _i ) is It is considered to be an outlier. Thus, by setting the values C and k, equation (1) can be used to determine the ith past transaction cost that meets the conditions for being an outlier.

  The exclusion unit 863 is configured to determine, from among the plurality of past transaction costs, one or more past transaction costs that meet the outlier condition.

  The setting unit 864 is configured to determine the predetermined limit for the user as the maximum value from the plurality of past transaction costs excluding one or more past transaction costs that satisfy the outlier condition.

  In one embodiment, clustering classification includes classification obtained after a predetermined amount of big data has undergone clustering processing.

  FIG. 9 is a functional diagram illustrating one embodiment of a programmed computer system for performing temporary transaction restriction generation. As is apparent, other computer system architectures and configurations can also be used to generate temporary transaction limits. Computer system 900, which includes various subsystems as described below, includes at least one microprocessor subsystem (also called a processor or central processing unit (CPU)) 902. For example, processor 902 may be implemented by a single chip processor or by multiple processors. In some embodiments, processor 902 is a general purpose digital processor that controls the operation of computer system 900. Using the instructions retrieved from memory 910, processor 902 controls the reception and manipulation of input data, as well as the output and display of data to an output device (e.g., display 918). In some embodiments, processor 902 is used to include and / or generate temporary transaction limits.

  Processor 902 is bi-directionally connected to memory 910, which is a first primary storage, typically random access memory (RAM), and a second primary storage area, typically read only memory (ROM). And can be included. As is well known in the art, primary storage can be used as a general purpose storage area and as a scratch pad memory, and can also be used to store input data and processed data. Primary storage may also store programming instructions and data in the form of data objects and text objects, in addition to other data and instructions for processes originating on processor 902. Again, as is well known in the art, primary storage is typically the basic operating instructions, program code, data, and so forth that processor 902 uses to perform its functions (eg, programmed instructions to name a few). Contains objects For example, memory 910 may include any suitable computer readable storage medium described below, depending on, for example, whether the data access needs to be bi-directional or uni-directional. For example, the processor 902 can also retrieve frequently needed data directly and very quickly and store it in a cache memory (not shown).

  Removable mass storage device 912 provides additional data storage capacity for computer system 900 and is connected to processor 902 in either a bi-directional (read / write) or uni-directional (read only) manner. For example, storage 912 may also include computer readable media, such as magnetic tape, flash memory, PC cards, portable mass storage devices, holographic storage devices, and other storage devices. For example, fixed mass storage 920 can also provide additional data storage capacity. The most common example of mass storage 920 is a hard disk drive. Mass storage 912 and 920 generally store such things as additional programming instructions and data not normally actively used by processor 902. It will be appreciated that the information retained in the mass storage 912 and 920 can be incorporated standardly into part of the memory 910 (e.g. RAM) as virtual memory, if desired.

  The bus 914 can also be used to enable access to other subsystems and devices in addition to enabling the processor 902 to access the storage subsystem. As shown in the figure, these include the display 918, the network interface 916, the keyboard 904, and the pointing device 908, as well as auxiliary input / output device interfaces, sound cards, speakers, and other subsystems as needed. It can be mentioned. For example, pointing device 908 may be a mouse, a stylus, a trackball, or a tablet and is useful for interacting with a graphical user interface.

  Network interface 916 allows processor 902 to be connected to another computer, computer network, or telecommunications network using a network connection as shown in the figure. For example, through the network interface 916, the processor 902 may receive information (eg, data objects or program instructions) from another network or output information to another network in the course of performing a method / process step Can. Information is often expressed as a series of instructions that are executed on a processor, and can be received from another network or output to another network. Implemented by an interface card or similar device and processor 902 to connect computer system 900 to an external network and to transfer data according to a standard protocol (eg, implemented / implemented on processor 902) Appropriate software can be used. For example, the various process embodiments disclosed herein may execute on processor 902, or cooperate with a remote processor that shares a portion of the process with the Internet, an intranet network, or a local area network. Etc. can be implemented in a network. An additional mass storage device (not shown) can also be connected to processor 902 through network interface 916.

  In cooperation with computer system 900, an auxiliary input / output device interface (not shown) can be used. The auxiliary input / output device interface is such that the processor 902 has a microphone, a touch-sensitive display, a transducer card reader, a tape reader, a voice or handwriting recognition device, a biometric reader, a camera, a portable mass storage device, and other computers. A general purpose as well as a dedicated interface may be included, which allows sending data to the device of, and more generally receiving data from these other devices.

  The systems, devices, modules or units described according to the above embodiments may be realized by a computer chip or entity or by a computer holding specific functions.

  For convenience of explanation, when describing the above devices, each function is described as a separate unit. Of course, when putting the present application into practice, the functionality of the various units may be implemented with the same or multiple software and / or hardware configurations.

  As can be seen through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized with the help of software and the necessary common hardware platform in addition thereto. . Based on such an understanding, the essence of the technical method of the present application, or the part contributing to the prior art can be embodied in the form of a software product. The computer software product may be stored in a storage medium such as a ROM / RAM, a magnetic disk or an optical disc, and the processes described by the various embodiments of the present application or particular parts of those embodiments , A server, or a network device).

  The present application may be used in many general purpose or special purpose computer systems or configurations. Examples include personal computers, servers, handheld devices or portable devices, tablet devices, multiprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, any of the above systems Or a distributed computing environment that includes the devices.

  The present application may be described in the general context of computer-executable commands, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., for performing particular tasks or implement particular abstract data types. The present application may also be performed within a distributed computing environment, where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located on storage media at a local or remote computer, including the storage device.

  Although the present application has been described through embodiments, one skilled in the art will appreciate that the present application has many variations and variations that do not depart from the spirit of the present application.

に基づいて決定され、ここで、ｉは、ｉ番目の取引を示し、ｄ ₁ 、ｄ ₂ 、……、ｄ _n は、ランク付けされた複数の過去取引費用を含み、Ｃは、閾値に設定され、ｋは、事前設定距離に設定される、システム。

適用例６：
適用例１に記載のシステムであって、
前記複数のユーザに関係付けられた前記過去取引データから、前記複数のクラスタ化分類を決定することは、前記複数のユーザに関係付けられた前記過去取引データの少なくとも一部分に対し、１つ以上の尺度にしたがってクラスタリングを実施することを含む、システム。

適用例７：
適用例６に記載のシステムであって、
前記１つ以上の尺度は、最新の消費、消費の頻度、および消費金額のうちの１つ以上を含む、システム。

適用例８：
適用例１に記載のシステムであって、
少なくとも部分的に、前記ユーザに関係付けられた前記過去取引データに基づいて、前記複数のクラスタ化分類から前記一つのクラスタ化分類を決定することは、
前記ユーザに関係付けられた前記過去取引データを、１つ以上の尺度に対応する値に変換することと、
前記１つ以上の尺度に対応する前記値を、前記クラスタ化分類に関係付けられた前記１つ以上の尺度に対応する値範囲と比較することと、
を含む、システム。

適用例９：
適用例１に記載のシステムであって、
前記一時的な取引制限は、前記あらかじめ定められた制限と前記動的クオータとの和を含む、システム。

適用例１０：
方法であって、
取引リクエストメッセージを受信することであって、前記取引リクエストメッセージは、総取引費用と、ユーザに関係付けられた識別情報とを含む、取引リクエストメッセージの受信と、
前記総取引費用があらかじめ定められた制限を上回ることを決定することと、
前記総取引費用が前記あらかじめ定められた制限を上回るとの決定を受けて、複数のユーザに関係付けられた過去取引データを取り出すことと、
前記複数のユーザに関係付けられた前記過去取引データから、複数のクラスタ化分類を決定することと、
１つ以上のプロセッサを使用し、少なくとも部分的に、前記ユーザに関係付けられた過去取引データに基づいて、前記複数のクラスタ化分類から一つのクラスタ化分類を決定することと、
あらかじめ定められたマッピングルールを使用して、前記ユーザのための前記クラスタ化分類に対応する動的クオータを決定することと、
少なくとも部分的に、前記総取引費用を一時的な取引制限と比較することに基づいて、前記取引リクエストメッセージが承認されるかどうかを決定することであって、前記一時的な取引制限は、前記あらかじめ定められた制限と前記動的クオータとの組み合わせを含む、ことと、
を備える方法。

適用例１１：
適用例１に記載の方法であって、
前記あらかじめ定められた制限は、ユーザ設定される、方法。

適用例１２：
適用例１に記載の方法であって、さらに、
前記あらかじめ定められた制限を決定することを備える方法。

適用例１３：
適用例１２に記載の方法であって、
前記ユーザに関係付けられた前記過去取引データは、前記ユーザに関係付けられた複数の過去取引費用を含み、前記あらかじめ定められた制限を決定することは、
前記ユーザに関係付けられた前記複数の過去取引費用をランク付けすることと、
外れ値条件を決定することと、
前記複数の過去取引費用から、前記外れ値条件を満たす１つ以上の過去取引費用を決定することと、
少なくとも部分的に、前記外れ値条件を満たす前記１つ以上の過去取引費用を除外した前記複数の過去取引費用からの最大値に基づいて、前記ユーザに関係付けられた前記あらかじめ定められた制限を決定することと、
を含む、方法。

適用例１４：
適用例１３に記載の方法であって、
前記外れ値条件は、少なくとも部分的に、

に基づいて決定され、ここで、ｉは、ｉ番目の取引を示し、ｄ ₁ 、ｄ ₂ 、……、ｄ _n は、ランク付けされた複数の過去取引費用を含み、Ｃは、閾値に設定され、ｋは、事前設定距離に設定される、方法。

適用例１５：
適用例１０に記載の方法であって、
前記複数のユーザに関係付けられた前記過去取引データから、前記複数のクラスタ化分類を決定することは、前記複数のユーザに関係付けられた前記過去取引データの少なくとも一部分に対し、１つ以上の尺度にしたがってクラスタリングを実施することを含む、方法。

適用例１６：
適用例１５に記載の方法であって、
前記１つ以上の尺度は、最新の消費、消費の頻度、および消費金額のうちの１つ以上を含む、方法。

適用例１７：
適用例１０に記載の方法であって、
少なくとも部分的に、前記ユーザに関係付けられた前記過去取引データに基づいて、前記複数のクラスタ化分類から前記一つのクラスタ化分類を決定することは、
前記ユーザに関係付けられた前記過去取引データを、１つ以上の尺度に対応する値に変換することと、
前記１つ以上の尺度に対応する前記値を、前記クラスタ化分類に関係付けられた前記１つ以上の尺度に対応する値範囲と比較することと、
を含む、方法。

適用例１８：
適用例１０に記載の方法であって、
前記一時的な取引制限は、前記あらかじめ定められた制限と前記動的クオータとの和を含む、方法。

適用例１９：
有体の非一過性のコンピュータ読み取り可能ストレージ媒体に盛り込まれたコンピュータプログラム製品であって、
取引リクエストメッセージを受信するためのコンピュータ命令であって、前記取引リクエストメッセージは、総取引費用と、ユーザに関係付けられた識別情報とを含む、コンピュータ命令と、
前記総取引費用があらかじめ定められた制限を上回ることを決定するためのコンピュータ命令と、
前記総取引費用が前記あらかじめ定められた制限を上回るとの決定を受けて、複数のユーザに関係付けられた過去取引データを取り出すためのコンピュータ命令と、
前記複数のユーザに関係付けられた前記過去取引データから、複数のクラスタ化分類を決定するためのコンピュータ命令と、
少なくとも部分的に、前記ユーザに関係付けられた過去取引データに基づいて、前記複数のクラスタ化分類から一つのクラスタ化分類を決定するためのコンピュータ命令と、
あらかじめ定められたマッピングルールを使用して、前記ユーザのための前記クラスタ化分類に対応する動的クオータを決定するためのコンピュータ命令と、
少なくとも部分的に、前記総取引費用を一時的な取引制限と比較することに基づいて、前記取引リクエストメッセージが承認されるかどうかを決定するためのコンピュータ命令であって、前記一時的な取引制限は、前記あらかじめ定められた制限と前記動的クオータとの組み合わせを含む、コンピュータ命令と、
を備えるコンピュータプログラム製品。 Although the foregoing embodiments have been described in some detail for purposes of clarity of understanding, the invention is not limited to the details provided. There are numerous alternative ways of implementing the invention. The disclosed embodiments are illustrative and not restrictive.
The present invention can also be realized, for example, in the following manner.

Application Example 1:
A system,
A receiving unit for receiving a transaction request message, the transaction request message including a total transaction cost and identification information associated with the user;
Query unit,
Determining that the total transaction cost exceeds a predetermined limit;
A querying unit for retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
A search unit,
Determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
A search unit that determines a clustered classification from said plurality of clustered classifications based at least in part on past transaction data associated with said user;
A mapping unit for determining a dynamic quota corresponding to the clustered classification for the user using a predefined mapping rule;
A verification unit that determines whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit being: A verification unit comprising a combination of the predetermined limit and the dynamic quota;
System with

Application Example 2:
The system according to Application Example 1 is,
The predetermined limit is user configured.

Application example 3:
The system according to Application Example 1 is,
The system, wherein the inquiry unit is further configured to determine the predetermined limit.

Application example 4:
The system according to Application Example 3 is,
The past transaction data associated with the user includes a plurality of past transaction costs associated with the user,
Determining said predetermined limit is:
Ranking the plurality of past transaction costs associated with the user;
Determining an outlier condition,
Determining one or more past transaction costs that satisfy the outlier condition from the plurality of past transaction costs;
The predetermined limit associated with the user based at least in part on a maximum value from the plurality of past transaction costs excluding the one or more past transaction costs meeting the outlier condition. To decide and
Including the system.

Application Example 5:
The system according to Application Example 4 is,
Said outlier condition is at least partially:

, Where i denotes the i th trade, d ₁ , d ₂ , ..., d _n comprises a plurality of ranked past transaction costs, and C is set to a threshold And k is set to a preset distance, the system.

Application Example 6:
The system according to Application Example 1 is,
The determining of the plurality of clustered classifications from the past transaction data associated with the plurality of users may be performed on at least a portion of the past transaction data associated with the plurality of users. A system comprising performing clustering according to a measure.

Application example 7:
The system according to Application Example 6 is,
The system, wherein the one or more measures include one or more of the latest consumption, frequency of consumption, and consumption amount.

Application Example 8:
The system according to Application Example 1 is,
Determining the one clustered classification from the plurality of clustered classifications based at least in part on the past transaction data associated with the user,
Converting the past transaction data associated with the user into values corresponding to one or more measures;
Comparing the values corresponding to the one or more measures with a value range corresponding to the one or more measures associated with the clustered classification;
Including the system.

Application Example 9:
The system according to Application Example 1 is,
The system, wherein the temporary transaction limit comprises the sum of the predetermined limit and the dynamic quota.

Application Example 10:
Method,
Receiving a transaction request message, the transaction request message comprising a total transaction cost and identification information associated with the user;
Determining that the total transaction cost exceeds a predetermined limit;
Retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
Determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
Determining one clustered classification from the plurality of clustered classifications using one or more processors and based at least in part on past transaction data associated with the user;
Determining a dynamic quota corresponding to the clustered classification for the user using a predefined mapping rule;
Determining whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit comprising Including a combination of predetermined limits and the dynamic quota,
How to provide.

Application Example 11:
The method described in Application Example 1 is,
The method wherein the predetermined limit is user configured.

Application Example 12:
The method according to Application Example 1, further comprising
A method comprising determining the predetermined limit.

Application Example 13:
A method described in Application Example 12;
The past transaction data associated with the user includes a plurality of past transaction costs associated with the user, and determining the predetermined limit is:
Ranking the plurality of past transaction costs associated with the user;
Determining an outlier condition,
Determining one or more past transaction costs that satisfy the outlier condition from the plurality of past transaction costs;
The predetermined limit associated with the user based at least in part on a maximum value from the plurality of past transaction costs excluding the one or more past transaction costs meeting the outlier condition. To decide and
Method, including.

Application Example 14:
The method described in Application Example 13 is,
Said outlier condition is at least partially:

, Where i denotes the i th trade, d ₁ , d ₂ , ..., d _n comprises a plurality of ranked past transaction costs, and C is set to a threshold And k is set to a preset distance, method.

Application Example 15:
The method described in Application Example 10, wherein
The determining of the plurality of clustered classifications from the past transaction data associated with the plurality of users may be performed on at least a portion of the past transaction data associated with the plurality of users. A method comprising performing clustering according to a measure.

Application Example 16:
The method described in Application Example 15 is,
The one or more measures include one or more of the latest consumption, frequency of consumption, and amount of consumption.

Application Example 17:
The method described in Application Example 10, wherein
Determining the one clustered classification from the plurality of clustered classifications based at least in part on the past transaction data associated with the user,
Converting the past transaction data associated with the user into values corresponding to one or more measures;
Comparing the values corresponding to the one or more measures with a value range corresponding to the one or more measures associated with the clustered classification;
Method, including.

Application Example 18:
The method described in Application Example 10, wherein
The temporary transaction limit comprises the sum of the predetermined limit and the dynamic quota.

Application Example 19:
A computer program product embodied in a tangible, non-transitory computer readable storage medium, comprising:
Computer instructions for receiving a transaction request message, the transaction request message including a total transaction cost and identification information associated with the user;
Computer instructions for determining that the total transaction cost exceeds a predetermined limit;
Computer instructions for retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
Computer instructions for determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
Computer instructions for determining a clustered classification from the plurality of clustered classifications based at least in part on past transaction data associated with the user;
Computer instructions for determining a dynamic quota corresponding to the clustered classification for the user using a predefined mapping rule;
A computer instruction for determining whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit Computer instructions, including a combination of said predetermined limits and said dynamic quotas;
A computer program product comprising:

Claims (19)

A system,
A generation unit configured to generate a classification unit used to select among a plurality of clustered classifications, wherein the classification unit is generated based at least in part on past data of the user, and A generation unit configured to perform classification selection from among the plurality of clustered classifications based at least in part on the associated input;
A receiving unit for receiving a transaction request message, the transaction request message including a total transaction cost and identification information associated with the user;
Query unit,
Determining that the total transaction cost exceeds a predetermined limit;
A querying unit for retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
A search unit,
Determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
Selecting one clustered classification from the plurality of clustered classifications based at least in part on past transaction data associated with the user, wherein selection of the one clustered classification is performed on the historical data of the user A search unit comprising applying a user score based on 基 づ く to the classifier to determine a clustered classification to be selected ;
A mapping unit for determining a dynamic quota corresponding to the clustered classification for the user using a predefined mapping rule;
A verification unit that determines whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit being: A verification unit comprising a combination of the predetermined limit and the dynamic quota;
System with The system according to claim 1, wherein
The predetermined limit is user configured. The system according to claim 1, wherein
The system, wherein the inquiry unit is further configured to determine the predetermined limit. The system according to claim 3, wherein
The past transaction data associated with the user includes a plurality of past transaction costs associated with the user,
Determining said predetermined limit is:
Ranking the plurality of past transaction costs associated with the user;
Determining an outlier condition,
Determining one or more past transaction costs that satisfy the outlier condition from the plurality of past transaction costs;
The predetermined limit associated with the user based at least in part on a maximum value from the plurality of past transaction costs excluding the one or more past transaction costs meeting the outlier condition. To decide and
Including the system. The system according to claim 4, wherein
Said outlier condition is at least partially:

, Where i denotes the i th trade, d ₁ , d ₂ , ..., d _n comprises a plurality of ranked past transaction costs, and C is set to a threshold And k is set to a preset distance, the system. The system according to claim 1, wherein
The determining of the plurality of clustered classifications from the past transaction data associated with the plurality of users may be performed on at least a portion of the past transaction data associated with the plurality of users. A system comprising performing clustering according to a measure. The system according to claim 6, wherein
The system, wherein the one or more measures include one or more of the latest consumption, frequency of consumption, and consumption amount. The system according to claim 1, wherein
Determining the one clustered classification from the plurality of clustered classifications based at least in part on the past transaction data associated with the user,
Converting the past transaction data associated with the user into values corresponding to one or more measures;
Comparing the values corresponding to the one or more measures with a value range corresponding to the one or more measures associated with the clustered classification;
Including the system. The system according to claim 1, wherein
The system, wherein the temporary transaction limit comprises the sum of the predetermined limit and the dynamic quota. Method,
Generating a classification unit used for selection among a plurality of clustered classifications, wherein the classification unit is generated based at least in part on past data of the user and is associated with past user data Generating a classification unit configured to perform classification selection from among the plurality of clustered classifications based at least in part on the determined input;
Receiving a transaction request message, the transaction request message comprising a total transaction cost and identification information associated with the user;
Determining that the total transaction cost exceeds a predetermined limit;
Retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
Determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
Using one or more processors, at least in part, based on past transaction data associated with the user, the method comprising: selecting one of the clustering classification from the plurality of clustering classification, the one Selection of one clustered classification, including applying a user score based on the past data of the user to the classification unit to determine a selected clustered classification ;
Determining a dynamic quota corresponding to the clustered classification for the user using a predefined mapping rule;
Determining whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit comprising Including a combination of predetermined limits and the dynamic quota,
How to provide. The method according to claim 10, wherein
The method wherein the predetermined limit is user configured. 11. The method of claim 10, further comprising:
A method comprising determining the predetermined limit. The method according to claim 12, wherein
The past transaction data associated with the user includes a plurality of past transaction costs associated with the user, and determining the predetermined limit is:
Ranking the plurality of past transaction costs associated with the user;
Determining an outlier condition,
Determining one or more past transaction costs that satisfy the outlier condition from the plurality of past transaction costs;
The predetermined limit associated with the user based at least in part on a maximum value from the plurality of past transaction costs excluding the one or more past transaction costs meeting the outlier condition. To decide and
Method, including. The method according to claim 13, wherein
Said outlier condition is at least partially:

, Where i denotes the i th trade, d ₁ , d ₂ , ..., d _n comprises a plurality of ranked past transaction costs, and C is set to a threshold And k is set to a preset distance, method. The method according to claim 10, wherein
The determining of the plurality of clustered classifications from the past transaction data associated with the plurality of users may be performed on at least a portion of the past transaction data associated with the plurality of users. A method comprising performing clustering according to a measure. The method according to claim 15, wherein
The one or more measures include one or more of the latest consumption, frequency of consumption, and amount of consumption. The method according to claim 10, wherein
Determining the one clustered classification from the plurality of clustered classifications based at least in part on the past transaction data associated with the user,
Converting the past transaction data associated with the user into values corresponding to one or more measures;
Comparing the values corresponding to the one or more measures with a value range corresponding to the one or more measures associated with the clustered classification;
Method, including. The method according to claim 10, wherein
The temporary transaction limit comprises the sum of the predetermined limit and the dynamic quota. A computer program, using a computer,
A function of generating a classification unit used for selecting among a plurality of clustered classifications, wherein the classification unit is generated based at least in part on past data of a user and is associated with past user data A function configured to perform a classification selection from among the plurality of clustered classifications based at least in part on the determined input;
A function of receiving a transaction request message, the transaction request message including a total transaction cost and identification information associated with the user;
A function of determining that the total transaction cost exceeds a predetermined limit;
A function of retrieving past transaction data associated with a plurality of users in response to the determination that the total transaction cost exceeds the predetermined limit;
A function of determining a plurality of clustered classifications from the past transaction data associated with the plurality of users;
A function of determining a clustered classification from said plurality of clustered classifications based at least in part on past transaction data associated with said user, the selection of said one clustered classification being: Applying a user score based on the user's past data to the classifier to determine the selected clustered classification ;
A function of determining a dynamic quota corresponding to the clustered classification for the user using a predetermined mapping rule;
A function of determining whether the transaction request message is approved based at least in part on comparing the total transaction cost to a temporary transaction limit, the temporary transaction limit comprising: A function, including a combination of predetermined limits and said dynamic quotas;
A computer program to make it happen.

Images